Cumsum in Python: Running Balances with itertools.accumulate

Start with a small, explicit task

Start a fictional token ledger at 100 and apply changes of -15, +30 and -10. Preserve the balance after each change, including the opening state. These are integer teaching tokens rather than money, and the example requires no NumPy or pandas installation.

Run the complete example

This original example was checked with Python 3.12.14. Save it as python_cumulative_sum.py and run it with Python 3.12. It uses only the standard library. The assertions check the stated results.

from itertools import accumulate
opening = 100
changes = [-15, 30, -10]
balances = list(accumulate(changes, initial=opening))
assert balances == [100, 85, 115, 105]
assert len(balances) == len(changes) + 1
assert balances[-1] == opening + sum(changes)
assert list(accumulate([], initial=opening)) == [100]
assert list(accumulate([])) == []
print(balances)
print("change total:", sum(changes))
print("closing balance:", balances[-1])

Explain the result

The sequence begins with 100 before any change. Subtracting 15 gives 85, adding 30 gives 115 and subtracting 10 gives 105. There is one more balance than change because the initial state is retained. If a report needs one row per transaction, deliberately align transaction rows with balances[1:] rather than silently shifting the numbers.

sum(changes) returns 5, which is the net change and not the closing balance. Adding the opening state produces 105. The empty checks distinguish no changes with an opening state from no changes without one. Materializing the iterator is useful for this small report; a streaming consumer can process running values directly.

Avoid these mistakes

Do not insert the opening balance both into the input changes and into initial, which would count it twice. Choose numeric types appropriate to your own domain; this exact integer exercise does not validate floating-point financial arithmetic or a third-party library’s missing-value behavior.